The Nineteenth Byte

Destructible Lemon

12:00 AM

@ATaco ? what?

ATaco

I once moved my entire folder structure into a sub folder.

Destructible Lemon

hah

ATaco

I had to re-install a-ta.co when that happened, and I had no source control..

Downgoat

please help how to python

Destructible Lemon

what is pip3

Downgoat

12:03 AM

wait crap wrong chatroom

@DestructibleLemon pip but for python3

Destructible Lemon

what's wrong with pip

ATaco

But pip is also for python3.

Destructible Lemon

clearly need to install configparser

Downgoat

@DestructibleLemon look up like 10 lines

Maltysen

@Downgoat slither around, grasp prey with your sharp teeth, and kill them through constriction

duh :P

Downgoat

12:09 AM

> Successfully installed ConfigParser-3.5.0

Destructible Lemon

install it better

Christopher 2EZ 4RTZ

@Downgoat at this point you do not deserve linux

Destructible Lemon

1:08 AM

I found some anti-intellectualism

(the quote in the question not the question)

Anonymous

2:34 AM

@Downgoat The issue, as pointed out in the other chatroom (which I'm saying here for others' benefits), is that the mysql package on PyPI is broken and not well-maintained. I'm looking into the terms of use to see if I can legally upload a working version to PyPI.

Pavel

@Mego The community python connector is licensed under the GPL, so probably.

ATaco

You can now call a function with a splat without parenthesis in Funky, eg. print ...t

quartata

3:14 AM

cmake pls

So basically

I installed the latest OS X version of CMake or whatever

and of course it sucks

first off the actual tools are hidden inside an application bundle because lul

secondly it doesn't actually work with GCC

>mfw correctly identifies compiler as GCC

>mfw tries to pass in clang options

>mfw The C++ compiler "/usr/local/bin/g++" is not able to compile a simple test program

>mfw refuses to use any other compiler even when I take /usr/local/bin out of the path

>mfw CMake is getting called from emsdk so I can't pass it a -D option to get it to use something else

Pavel

3:28 AM

Huh, today is 100 years since the october revolution

quartata

OK I managed to make a bash wrapper that makes it work

I'm sure it'll piss me off in six months when I run cmake on something and it starts building it with clang

I'll be back in 8 hours once emscripten builds

which by the way I'm having to build from source because this release has an emar that just flat out doesn't work on OS X

bye

aditsu

@Pavel before I post this challenge and somebody answers it in C#, could you give me the code for a trivial GUI app? ALL the information I found on the net involves designing a window in Visual Studio.. also, zsh: command not found: dotnet

Pavel

@aditsu Have you installed .net core

quartata

@aditsu you need .net core

ninja'd

Sherlock9

@Pavel Darn Julian calendar, putting the October Revolution in November

aditsu

3:34 AM

@Pavel not sure how to go about doing that, will search..

Pavel

K

@aditsu Apparantly Core does not support GUI. You could use Mono with GTK#, which has the same APIs as Windows Forms.

aditsu

ok, still need the source code for the program though..

Pavel

@aditsu Remember what I said about no small C# projects? I create this simple app with the form designer and came out of it with 4 .cs files.

quartata

> form designer

Pavel

Presumably there's a way to do it simply with one file but really when it comes to forms you're not supposed to do it

@quartata The Visual Studio one

quartata

3:43 AM

problem_spotted.gif

aditsu

right, if you can't minimize it, just zip them up

Pavel

Well IDK any other way to make a gui

@aditsu drive.google.com/file/d/13Py8KH4EDReKNiZOc4wUmyWf0_ToIh-g/…

aditsu

thanks

btw, in java you can simply do public class X extends javafx.application.Application{public void start(javafx.stage.Stage s){s.show();}} -> bam, window

Pavel

@aditsu I'm sure C# supports that but I have literally never made a GUI application in my life

aditsu

now I have to figure out how to compile your project :p

I'll try msbuild ..maybe

quartata

3:53 AM

oh wonderful

turns out it's still broken in the incoming version of emscripten

so that was for nothing

aditsu

ooh, I think I managed to compile it just with mcs

Pavel

Nice

The dotnet core compiler is super intelligent, dotnet build (or dotnet run). The tools make the needless complexity bearable. Unfortunatly, you don't get GUIs.

Anonymous

I hate that Microsoft only wants VS build tools to be installed on C:. That's utterly stupid for the users that have multiple drives.

ATaco

Just swap your drive order, EZPZ

Anonymous

@ATaco Not so EZPZ when Windows is literally hardcoded to expect system files on C:, and for C: to be the boot drive

ATaco

4:05 AM

I know, I was joking.

I prefer the unix / over C:\ anyway :P

Anonymous

At this point my solution is to copy my working directory over to my Linux server and compile it there

aditsu

@Pavel well, I don't have "dotnet" but I was able to run your code

I stripped it down to this: using System.Windows.Forms;static class X{static void Main(){Application.Run(new Form());}}

Pavel

Nice

@aditsu The static class can be shortened to class. Main can become static void Main()=>Application.Run(new Form());

@aditsu You don't need main?

aditsu

@Pavel nope, javafx application

Pavel

Ah

aditsu

4:18 AM

you can make it shorter using AWT and including a main method, but it doesn't behave nicely..

@Pavel I don't think "=>" makes it shorter..

ATaco

Funky's syntax is fine.

Pavel

@aditsu I suppose it doesn't. It only makes it shorter when there's no return statement.

@ATaco Wat the even

Anonymous

4:44 AM

1
0/1
0

Pavel

@ATaco You have to explain what's going on there

ATaco

The tokenizer wants to allow either func or function, so it sees func as one part and tion as an optional part, and Funky lets whitespace live between tokens.

Pavel

@ATaco If tokens are regexes, the tion token should be (?<=func)tion

Actually, if tokens are regexes, func/tion should be one token: func(?:tion)?

ATaco

It should be one token, but it isn't.

Anonymous

@ATaco How are function definitions supposed to look?

Anonymous

4:54 AM

Like say I wanted to write def f(): print("foo") in (readable) Funky

ATaco

function a(){} func a(){} function a() b function a x{} (a, b)=>{} a=>{}

function f(){
	print("foo")
}

Anonymous

This is why loose tokenizing is a bad idea :P

ATaco

The second func there is dropped :P

Pavel

@Mego ಠ_ಠ

ATaco

It's written such that anyone trying to write sane code won't ever notice it.

Anonymous

4:59 AM

@Pavel Point those at the one who made the language, not me

ATaco

Shhhh, Funky is beautiful.

Look at how pretty this answer is!

Pavel

5:15 AM

@Mego god damnit the first func can be dropped entirely

Tion is enough for function declaration

ATaco

That's not actually defining a function.

Pavel

@ATaco Well what is it doing

ATaco

Pavel

@ATaco tio.run/##SyvNy678/78kMz9PIVFDU6FaoaAoM69EQyktP19JU6GWCyj4/z8A

ATaco

Does nothing with tion, tries to call nothing, then tries to call that result with a table that contains print("foo") which is parsed.

So most of it is ignored.

Pavel

5:22 AM

Seriously tho you should fix function token behavior

I shouldn't be able to define a function with tion

ATaco

You can't, that's not defining a function.

If you remove the final func tion(), then it still prints foo

Try it online!

Pavel

Oh

Shouldn't it error or something

Anonymous

@Pavel Errors are overrated

ATaco

There are many parts of this that should logically error, but that's boring.

Anonymous

It's like you took the overeager type deduction schemes from JS and PHP and applied them to the whole language's syntax

ATaco

5:29 AM

That's a pretty good description.

Pavel

@Mego I'm trying to think of a language with a similar approach to the idea of NameError and I can't

Anonymous

You were so preoccupied with whether or not you could that you didn't stop to think if you should.

ATaco

Funky is by far my best made language, too.

Anonymous

That's frightening

ATaco

Well my last few languages are held together with sticky tape.

This is my first attempt at a C-like (Javascript-like) language, and I'd think it's pretty good for what it is.

function gcd (a, b) {
    while (b != 0) {
        c = a % b;
        a = b;
        b = c;
    };
    return a;
}

function A000938 (n) {
    n = n + 2;
    ans = 0;
    for (m = 2; m <= n; ++m) {
        for (k = 2; k <= n; ++k) {
            ans = ans + (((n - k) + 1) * ((n - m) + 1) * gcd(k - 1, m - 1));
        }
    };
    ans = (2 * ans) - ((n*n)*(((n*n) - 1)/6));
    return ans;
}

See? That actually looks like code!

Pavel

5:38 AM

@Mego I invite you to look at the source code for RProgN2

ATaco

(It's also a polyglot of Javascript)

github.com/TehFlaminTaco/RProgN-2/tree/master/source/src/rprogn

Good luck

Anonymous

@ATaco Do integers not decay to booleans for conditionals?

ATaco

What do you mean?

Oh, they do, but that was for the OEIS challenge and bytecount manipulation is important.

Anonymous

Ah

Anonymous

So while(b) is an acceptable alternative for while(b != 0), but for challenge-specific reasons you chose the latter

ATaco

5:41 AM

function gcd (a, b) {
    while b{
        c = a % b;
        a = b;
        b = c;
    };
    return a;
}

I didn't write that answer

5

A: One OEIS after another

232. Funky, 326 bytes, A000938 function gcd (a, b) { while (b != 0) { c = a % b; a = b; b = c; }; return a; } function A000938 (n) { n = n + 2; ans = 0; for (m = 2; m <= n; ++m) { for (k = 2; k <= n; ++k) { ans = ans + (((n - k...

Anonymous

@Pavel I've seen the source, I think. That's the one with the million different classes for things that should be instances, right?

Pavel

@Mego yep

ATaco

RProgN1 is almost entirely 1 file.

Gotta have some contrast.

Pavel

All I want in life is so that the square button on the Android navbar lets you switch between chrome tabs, not just apps.

Anonymous

@Pavel Wish granted. You are now unable to ever use Android devices again, including emulated devices.

ATaco

5:46 AM

My least favourite thing about Funky is that operators are RTL.

Including Unary Operators.

Anonymous

@ATaco Why not merge two projects together and make them AnyFix?

ATaco

!a+b is equivalent to !(a+b) rather than (!a)+b

Pavel

@Mego what why

Anonymous

@Pavel Chaos

ATaco

If I had used prefix rather than infix it would have been infintiely better.

Pavel

5:49 AM

Still, the amount of times I hit square by accident...

Like literally just now ._.

Marcus Andrews

Any tips for this one?
print(sum(any(k+k in w for k in map(chr,range(65,91)))for w in input().upper().split()))

Counting number of words that have at least one double-letter (case doesn't matter)

double letter being "aa", "bb", ..., "zz"

shorten / ruby / etc?

Anonymous

@MarcusAndrews Are overlaps counted? e.g. is aaa 1 or 2?

Marcus Andrews

the word is all that's counted, as long as at least one such double letter exists

e.g. "balloon" is 1

"aaa" is also 1

Anonymous

Ah

Marcus Andrews

and the letters do have to be adjacent, e.g. aba = 0 but aab = 1

Checked your profile because I dug the penguin avatar -- and ps if you like math + programming, check out Project Euler

Anonymous

5:57 AM

import re;print(len(re.findall(r"(\w*(\w)\2\w*)",input(),re.I))) should work

Marcus Andrews

indeed it does

Anonymous

Marcus Andrews

:P

Anonymous

I started trying to solve that problem the hard way, and as I was typing out a regex to match to verify that I understood the problem, I realized I was being dense by not using regex

Anonymous

Also -2 bytes by removing the outer parens in the regex string and changing \2 to \1

Marcus Andrews

6:00 AM

I had tried a regex too but it was too long

Anonymous

import re;print(len(re.findall(r"\w*(\w)\1\w*",input(),re.I)))

Marcus Andrews

So this inherently searches words since it's not including whitespace, any letter in any amount, followed by a particular letter group capture, followed by a repeat of that letter, followed by any more chars in any amount

Anonymous

Yep

Anonymous

re.I is a flag that tells it to do a case-insentitive search

Anonymous

It's short for re.IGNORECASE

Marcus Andrews

6:03 AM

So even if I had put something like A-Za-z in there it's going to ignore the case anyway?

err, A-Z, let's say

ATaco

RProgN 2, 12 bytes û{S#eá+1m}r+

Marcus Andrews

and the text was all lower

Anonymous

Not quite. It ignores the case of the text, so A-Z is equivalent to A-Za-z

Marcus Andrews

import re
print(re.findall("[A-Z]", "lowercase text")) # []
print(re.findall("[A-Z]", "lowercase text", re.I)) # ['l', 'o', 'w', 'e', 'r', 'c', 'a', 's', 'e', 't', 'e', 'x', 't']

seems to work

yea, that is what I meant (I explained myself poorly)

Qwerp-Derp

6:15 AM

Hello all!

ATaco

oh, heya Qwerp.

Qwerp-Derp

I think I have a challenge idea

So I have a line of an even number of skittles, which are red and green. Half of the skittles are red and half are green. I can swap two skittles around, but every time I swap the skittles around I have to create a new permutation of skittles. What is the size n of the biggest swap I have to do, where n is the amount of skittles in between the two swapped skittles?

For 4 skittles, the result is 1, since I can go RRGG -> RGRG -> GRRG -> GRGR -> RGGR -> GGRR, with the final swap being the largest, with the two skittles being 1 skittle apart.

ATaco

Oh, that’s kind of interesting

I’d have no idea how to solve it

Wait, I know a bad way

Anonymous

@ATaco O((n!)!)?

ATaco

6:35 AM

I’m not sure the complexity

Anonymous

Pure brute force would brute force each step (O(n!)). It should come out to O((n!)!) total.

Marcus Andrews

what's the end goal? reversing the skittles?

Qwerp-Derp

@MarcusAndrews Not necessarily, you have to go through all permutations

I didn't make that clear, whoops :P

Marcus Andrews

6:59 AM

do you always have to pick the smallest possible swap for a given move?

nvm

Stewie Griffin

@Fatalize strictly speaking, "Shouldn't be too hard" = "Someone should be able to do it"..

Marcus Andrews

@Qwerp-Derp Why wouldn't the answer for 4 be 2? e.g. ['RRGG', 'GRGR', 'RGGR', 'RGRG', 'GRRG', 'GGRR'], the first move counted as 2 if we swap positions 0 and 3? (2 skittles in between)

Qwerp-Derp

7:29 AM

@MarcusAndrews It should be the minimal maximum, so it's sort of a minimax algorithm

Marcus Andrews

or are you asking for min of the max swap dists over all permutation paths? (i.e. no matter which path we take, all paths have at least one max swap of size X somewhere)

aagh ninja'd

Qwerp-Derp

@MarcusAndrews Bingo

New Main Posts

9:18 AM

3

Q: Seqindignot sequence

Title is made up, from 'Sequence Index Digit Not'. Challenge: Given an integer n which is >= 0, output the n'th number of the following sequence. Here are the first 50 items, with it's (0-indexed) index above it: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 3...

New Main Posts

9:33 AM

-2

Q: Find the second largest number in an array

Inspired by THIS question I present to you a coding challenge. The Challenge Find the second largest number in an array. Rules You cannot sort the array You cannot use any array specific functions (Like Math.max() in JS) Test Cases [-1, -7, -8, -9, -2] // -2 [-1, -7, 8, 9, -2] // 8 [1, 7...

popularity-contest programming-puzzle

New Sandboxed Posts

10:33 AM

0

A: Sandbox for Proposed Challenges

The lowest initial numbers in a Fibonacci-like sequence Given a positive integer input N, output the two non-negative numbers, a and b with the lowest possible mean value that will result in the number N being part of the recurring relation sequence: f(0) = a f(1) = b f(n) = f(n-2)+f(n-1) In ...

Erik the Outgolfer

12:23 PM

@Fatalize would you pay me if I say "hey, why not use prolog since it's way easier to and the invocation command will be significantly shorter"? ;-)

@Uriel the key is to not remember the builtins, but use the docs instead

only the syntax is worth remembering imo

although remembering the commonly-used builtins will make golfing significantly faster, you will remember them anyways over time

Adám

@EriktheOutgolfer Or use a language where the built-ins are mnemonic…

Erik the Outgolfer

@Adám um, that's very, very unlikely for all builtins, if not outright impossible...

although apl does try significantly to do that

but not all are perfectly mnemonic...

Adám

12:42 PM

@EriktheOutgolfer Not perfect, but pretty good, especially if one is made aware of how they are mnemonic. Do you have an example in mind of an APL symbol for which is neither a standard mathematical symbol, nor has a connection between its shape and what it does?

Erik the Outgolfer

monadic + :p

@Adám it does complex conjugate, for which I don't see any relation to +

or ¨ which is the each operator

Adám

@EriktheOutgolfer Well, + is obviously a companion to -. So while - negates both the real and the imaginary parts, + only negates the imaginary part. Clearly, there could be a third monadic primitive which only negates the real part, but this is very rarely needed (I think) and can trivially be done with +∘- or -∘+. This is similar to how there is no secondary-diagonal transpose, as it can be trivially done with ⌽∘⍉∘⌽.

@EriktheOutgolfer What? ¨ has one dot for each item (minimum two to be meaningful in normal circumstances).

Erik the Outgolfer

@Adám another example could be ⍎ (eval) and ⍕ (str)

or ⍝ for comments :p

Adám

@EriktheOutgolfer ⊥ is a base. It decodes from certain representation to normal numbers. ⍎ decodes from string representation to normal numbers (or whatever result).

@EriktheOutgolfer ⊤ and ⍕ are of course the inverses (as well as they can be inverses).

Erik the Outgolfer

yet another example would be ⌷ I guess?

(index)

Adám

12:53 PM

@EriktheOutgolfer That's a lightbulb or a light bulb filament to enlighten you, just like ideas are represented by light bulbs over cartoon characters' heads. (Yeah, the font here is not good, but in a proper APL font it looks much more like that.)

@EriktheOutgolfer That one is just the indexing square brackets smooshed together: [] → ⌷

Erik the Outgolfer

in the apl385 font the light bulb looks just like 0x7f :p

Adám

@EriktheOutgolfer It really is, btw. If you type [ and then go back to that character in overwrite mode and type ] you will get a real ⌷.

@EriktheOutgolfer A house‽

Erik the Outgolfer

huh

oh wait

Fatalize

@EriktheOutgolfer I wouldn't pay you either way :p

Erik the Outgolfer

it used to be different o_o

Adám

12:57 PM

Erik the Outgolfer

@Adám yes that's what 0x7f () looked like in chrome in the past

Fatalize

isnt the circle not supposed to fuse with the outer line?

Adám

@Fatalize What, in a real light bulb or in the APL character?

Erik the Outgolfer

@Fatalize well, life isn't perfect :p

Adám

If you look at the photo of the real light bulb, you can see that from a certain angle, all the hoops would overlay each other and the total shape would basically be that of the properly rendered APL symbol.

Erik the Outgolfer

1:01 PM

I think it's supposed to show the enlightening, not the bulb construct :p

Fatalize

@Adám the APL character

Adám

@Fatalize No, that is a mistake in many non-APL fonts. The image I posted is from a widely used font made by an APL old-timer.

@Fatalize And this is IBM's official APL font:

Fatalize

TIL

Not sure if I like that better than the incorrect symbol though

Adám

This is APLX's official font:

Not to clutter too much, this last one is from the IBM System/360 typewriter font, which is also used by APL+:

HyperNeutrino

my CS class's teaching order: stuff not related to CS -> String output + input -> GUI design -> Numbers

user202729

1:12 PM

Numbers should come first IMO.

jaytea

The first one looks like a guy bending over

HyperNeutrino

I think so too (about the numbers)

user202729

So what have you learned recently about numbers?

jaytea

Perhaps a Kevin Spacey scene

HyperNeutrino

how to do like the most basic things like ++, --, +, -, *, /, etc

And what annoys me is that it doesn't mention that a++ evaluates as a and ++a evaluates as a + 1

user202729

1:14 PM

The key to write unparenthesized code: Always do -~ and ~- and assign instead of ++ and --.

HyperNeutrino

Then again, I don't think the teacher will be expecting us to do anything that makes that matter :P

Adám

@jaytea I see what you mean, it kind of has a nose.

HyperNeutrino

@user202729 +1 that is definitely good idea

user202729

@HyperNeutrino That is not supposed to cover in "simple" courses.

HyperNeutrino

true

user202729

1:15 PM

@HyperNeutrino That's your idea.

jaytea

More like a head-on view of a dude on his hands and knees

Aka just another day in the life of Kevin Spacey

Adám

@jaytea Maybe the dude is commenting upon your code.

HyperNeutrino

@user202729 huh?

Adám

A further informal APL learning session is taking place this evening at 18:30 UTC - you can join Adam at https://chat.stackexchange.com/rooms/52405/apl Reading along is open to everyone, but 20 StackExchange reputation points are required to contribute messages to the chat.

Tweeted by dyalogapl on November 8, 2017 at 12:28 PM

user202729

Oct 17 at 14:15, by user202729

@HyperNeutrino And what's with int[length * ~-length * ~-~-length / 6] and combine(nodes, old, (x, y) -> x ^ y)? Are you going to obfuscate the code?

HyperNeutrino

1:17 PM

oh lol right

jaytea

He's trying to get a closer look at my code because it's so small and well-golfed!!

haha

HyperNeutrino

next OEIS looks interesting

anyway must write history test now lol gtg o/

jaytea

cya, stay in school \o

user202729

And I, again, get disappointed by the solution to the planar map oeis sequence.

HyperNeutrino

1:33 PM

s/test/quiz/

that was a short quiz lol

@user202729 was it a formula copy ._.

user202729

Google classroom quiz?

HyperNeutrino

not for history, it was a paper quiz

user202729

Basically so.

HyperNeutrino

ah ok ._.

user202729

The next one is basically good, but that's because the oeis definition provide formula.

Except that it does so in the slowest way.

HyperNeutrino

1:35 PM

lol

user202729

(like, Fibonacci with recursion)

HyperNeutrino

ew lol

user202729

I may be wrong though.

idk what is the T function.

No problem, T is probably true.

but I find it weird that it runs in exactly 60.003 seconds on TIO, for both input 9 and 10.

(real time)

HyperNeutrino

hm interesting

aren't 9 and 10 both 0?

oh you were talking about [249]

user202729

Somehow it looks like that the TIO interpreter of picolisp will always run for exactly 60 seconds and then terminate with a trailing :. Weird.

@Dennis have a look?

HyperNeutrino

1:41 PM

interestingly enough the user time is quite short lol

user202729

and that the functions seems to work with larger values, probably with memoization.

> A self-complementary graph is one isomorphic to its complement.

So simple definition.

muddyfish

theguardian.com/technology/2017/nov/08/…

looks like more fun

user202729

So basically <generate all graph> <deduplicate by isomorphism> <check isomorphism with complement> <count>

There must be a better way.

HyperNeutrino

that's pretty much what I'd do and it sounds very slow xD

or use the subtraction formula given but I don't know why it works so I won't

user202729

> I don't know why it works so I won't

+100. (I don't have that many rep anyway)

HyperNeutrino

1:55 PM

:P

Uriel

@EriktheOutgolfer there's still the matter of knowing the general syntax/ structures. I can use 05ab1e even though I don't even remember the prime built in, but won't be able to write something in jelly beyond a 4 byter

Erik the Outgolfer

@Uriel stack-based languages are often very simple

HyperNeutrino

@Uriel I don't remember any builtins; I just open the docs and copy-paste :D

oh also you only have to generate all graphs with (n-1)!/2 edges for n nodes but I'm sure you already had that figured out

user202729

Actually n(n-1)/4.

Because number of vertices of a complete graph with n vertices is n(n-1)/2.

HyperNeutrino

oh nvm lol

Uriel

1:59 PM